Tech Workers in China Are Being Asked to Train AI Agents to Replace Them, Sparking Existential Questions
The landscape of work in China’s burgeoning tech sector is undergoing a profound and unsettling transformation. A growing number of tech professionals are reporting that their employers are instructing them to meticulously document their workflows and skills, not for knowledge transfer or training purposes, but to meticulously train artificial intelligence agents designed to ultimately replace them. This directive, amplified by a viral spoof tool that inadvertently captured the anxieties of the workforce, is igniting a wave of introspection among individuals who have historically been enthusiastic early adopters of technological advancements. The trend raises critical questions about the future of employment, the definition of human value in the workplace, and the ethical considerations surrounding AI’s integration into professional life.
The Genesis of "Colleague Skill" and its Unintended Resonance
The catalyst for much of this recent soul-searching emerged earlier this month with the sudden virality of a GitHub project named "Colleague Skill." Ostensibly a playful tool, it claimed to enable users to "distill" the skills, personality traits, and operational nuances of their colleagues, thereby creating AI agents capable of replicating their work. While initially conceived as a spoof, the project struck a deeply resonant chord within China’s tech community. Numerous professionals have since confided to MIT Technology Review that their superiors are actively encouraging them to codify their daily tasks and processes using AI agent tools such as OpenClaw or Claude Code.
The mechanics of "Colleague Skill" are deceptively simple. A user identifies a colleague whose duties they wish to emulate, providing basic profile information. The tool then proceeds to automatically import chat histories and documents from Lark and DingTalk, two widely used workplace communication and collaboration platforms in China. The output is a comprehensive, reusable manual detailing the targeted colleague’s responsibilities, including their unique idiosyncrasies and working styles, all curated for an AI agent to learn and replicate.
The creator of "Colleague Skill," Tianyi Zhou, an engineer at the Shanghai Artificial Intelligence Laboratory, stated in an interview with the Chinese outlet Southern Metropolis Daily that the project was indeed launched as a provocative stunt. He cited the increasing prevalence of AI-related layoffs and the growing corporate inclination to prompt employees to automate themselves as the primary motivations. Zhou did not respond to further requests for comment, leaving the interpretation of his intentions open.
A Mirror to Workplace Anxiety and Emerging Realities
The humorous aspect of "Colleague Skill" was not lost on internet users, many of whom jested about automating their colleagues before themselves. However, beneath the surface-level amusement lies a much more profound debate concerning worker dignity, individuality, and the potential dehumanization of labor in an era increasingly defined by artificial intelligence.
Amber Li, a 27-year-old tech worker in Shanghai, exemplifies this sentiment. After encountering "Colleague Skill" on social media, she used it as a personal experiment to recreate a former colleague. Within minutes, the tool generated a detailed file outlining that individual’s job functions. "It is surprisingly good," Li remarked. "It even captures the person’s little quirks, like how they react and their punctuation habits." Li found she could then utilize an AI agent as a virtual "coworker" capable of assisting with code debugging and providing instantaneous responses. The experience, however, was profoundly unsettling, described by Li as "uncanny and uncomfortable."
The increasing push from employers suggests that replacing human workers with AI agents could rapidly transition from a speculative concept to a workplace norm. The national fervor surrounding OpenClaw, a popular AI agent tool, has intensified this trend, with bosses across China actively urging their tech employees to explore and experiment with these advanced agents.
While AI agents are demonstrably capable of performing a wide array of tasks – from controlling computers and summarizing news to managing emails and booking reservations – on-the-ground reports from tech workers indicate that their practical utility in complex business environments remains somewhat limited. The process of asking employees to create detailed manuals of their day-to-day responsibilities, as facilitated by "Colleague Skill," is seen by some as a direct attempt to bridge this utility gap, extracting actionable knowledge that AI can effectively leverage.
The Corporate Rationale and the Employee Experience
Hancheng Cao, an assistant professor at Emory University specializing in the study of AI and work, posits that companies have strategic motivations beyond merely participating in a technological trend when they encourage employees to develop these "work blueprints." "Firms gain not only internal experience with the tools, but also richer data on employee know-how, workflows, and decision patterns," Cao explained. "That helps companies see which parts of work can be standardized or codified into systems, and which still depend on human judgment." This insight suggests a dual objective: to identify areas ripe for automation and to gain a deeper understanding of the non-codifiable elements of human expertise.
However, for the employees tasked with this process, the creation of AI agents or the blueprints for them can feel deeply alienating and dehumanizing. One software engineer, who requested anonymity due to concerns about job security, described the experience of training an AI on their workflow (using a tool other than "Colleague Skill") as profoundly reductive. The process, they felt, "flattened" their work into discrete modules, making them more easily replaceable. This sentiment is echoed on social media platforms. On Rednote, a user lamented, "a cold farewell can be turned into warm tokens," wryly suggesting that by first distilling their colleagues into tasks using "Colleague Skill," they might prolong their own employment.
Countermeasures and the Evolving Dialogue
The pervasive push for AI agent creation has also inadvertently spurred ingenious countermeasures. Koki Xu, a 26-year-old AI product manager in Beijing, found the idea of reducing a person to a mere skill set so objectionable that she developed an "anti-distillation" tool. Published on GitHub on April 4th, this tool, which took Xu approximately an hour to develop, is designed to actively sabotage the workflow documentation process for AI agents. Users can select from light, medium, or heavy sabotage modes, depending on the perceived level of managerial scrutiny. The agent then subtly rewrites the input material into generic, non-actionable language, thereby creating a less useful AI stand-in. A video demonstrating Xu’s project garnered over 5 million likes across various platforms, highlighting the widespread appeal of her resistance.
Xu shared with MIT Technology Review that she had been observing the "Colleague Skill" trend since its inception and that it prompted significant reflection on alienation, disempowerment, and the broader implications for labor. "I originally wanted to write an op-ed, but decided it would be more useful to make something that pushes back against it," she stated.
With undergraduate and master’s degrees in law, Xu also highlighted the emerging legal complexities surrounding this trend. While companies might assert ownership over work chat histories and materials created on company equipment, the "Colleague Skill" tool’s ability to capture elements of personality, tone, and judgment introduces significant ambiguity regarding intellectual property rights. She expressed a hope that "Colleague Skill" would stimulate a more robust discussion on how to safeguard workers’ dignity and identity in the AI era. "I believe it’s important to keep up with these trends so we (employees) can participate in shaping how they are used," Xu emphasized, noting her own active engagement with AI tools, including seven OpenClaw agents deployed across her personal and work devices.
For tech workers like Amber Li in Shanghai, the immediate threat of replacement remains somewhat distant, largely due to the current unreliability of AI tools and their persistent need for human oversight. "I don’t feel like my job is immediately at risk," Li admitted. "But I do feel that my value is being cheapened, and I don’t know what to do about it." This sentiment encapsulates the pervasive unease within the sector, where the abstract concept of AI replacing jobs is beginning to manifest as a tangible erosion of perceived worth, even without immediate job loss.
The Broader Implications: A Shifting Paradigm of Work
The trend of tech workers being tasked with creating AI agents to potentially replace themselves is not merely a localized phenomenon in China; it signals a broader global shift in the nature of work. As AI technologies mature, the demand for human labor in tasks that can be codified and automated is likely to diminish. This raises significant societal questions about income inequality, the need for reskilling and upskilling initiatives, and the potential for new forms of employment that leverage uniquely human capabilities such as creativity, critical thinking, and emotional intelligence.
The development of tools like "Colleague Skill" and its subsequent "anti-distillation" counterpoint highlight a growing tension between corporate efficiency goals and individual worker autonomy and dignity. The legal and ethical frameworks governing AI in the workplace are still in their nascent stages, and the current situation underscores the urgent need for thoughtful policy development and robust public discourse.
The current reality for many Chinese tech workers is one of profound uncertainty. While they are instrumental in developing the very tools that might render their roles obsolete, they are also grappling with the existential implications of their contributions. The future of work, as shaped by these AI-driven transformations, will undoubtedly require innovative solutions and a recalibration of how human value is recognized and rewarded in an increasingly automated world. The debate ignited by "Colleague Skill" is a critical early chapter in this ongoing narrative, forcing a confrontation with the complex ethical and practical challenges that lie ahead.
